What is New in Toga II 1.3x4?

[IWB]

best is to say how the average NPS of the programs is on that hardware.
for terrys machine we see it is 107 KNps.

Hmm, I dont know why this is of any interest, and I can not access my testcomp right now, but this is from my second comp which is about 5-10% slower then the machine Toga II 1.3x4 run the test on.

Starting position, 256 MB Hash, 1 Minute run:

rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq -

Engine: Toga II 1.3x4 64 (256 MB)
by Thomas Gaksch and Fabien Letouzey

12/30 0:01 +0.27 1.Nc3 Nf6 2.Nf3 Nc6 3.d4 d5 4.e3 Bg4
5.Bb5 e6 6.O-O Bd6 (1.540.722) 870

13/30 0:03 +0.14 1.Nc3 d5 2.d4 Nc6 3.Bf4 f6 4.e4 dxe4
5.Bb5 Bd7 6.Bc4 Bf5 7.Qh5+ g6 (3.401.967) 856

13/32 0:05 +0.22 1.d4 Nf6 2.Nf3 d5 3.Bf4 e6 4.e3 Bd6
5.Bd3 Bxf4 6.exf4 O-O 7.O-O Nc6 (4.817.434) 864

14/36 0:11 +0.25 1.d4 Nf6 2.Nf3 d5 3.Bf4 Bf5 4.e3 e6
5.Bb5+ c6 6.Bd3 Bxd3 7.Qxd3 Bd6
8.O-O O-O (9.725.780) 863

15/36 0:19 +0.20 1.d4 Nf6 2.Nf3 e6 3.Nc3 d5 4.Bg5 Nc6
5.e4 Be7 6.Bxf6 Bxf6 7.Bb5 O-O 8.O-O dxe4
9.Nxe4 (17.091.984) 862

15/51 0:37 +0.30 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Bxc6 dxc6
5.O-O Bg4 6.h3 Bxf3 7.Qxf3 Nf6 8.Nc3 Bc5
9.d3 (32.659.277) 875

best move: e2-e4 time: 1:00.750 min n/s: 885.305 CPU 100.0% n/s(1CPU): 885.305 nodes: 53.630.000

I consider this as resonable fast.

I tried some compiles from here: http://cito.atspace.com/
and the one I use is the fastest one on my comps.

Bye
Ingo

[Graham Banks]

best is to say how the average NPS of the programs is on that hardware.
for terrys machine we see it is 107 KNps.

Whilst this is important information, I don't think we should be discouraging hobbyists with slower hardware from indulging in their hobby or posting their results.

I've been in Terry's position in that regard and known how it feels.

Comparisons done between blitz and long time control ratings show that in the majority of cases there is little difference.
Of course, one gets higher quality games at longer time controls, but that's a different issue again.

Just my opinion.

Regards, Graham.

[nuff]

best is to say how the average NPS of the programs is on that hardware.
for terrys machine we see it is 107 KNps.

Whilst this is important information, I don't think we should be discouraging hobbyists with slower hardware from indulging in their hobby or posting their results.

Regards, Graham.

There are people with all sorts of hardware here so Terry should not feel ashamed. It is best to provide benchmark results as well as the settings/pgns produced so that any mitigating factor can be deciphered in the result. Can you imagine if the SSDF or CCRL did not state what hardware they were using because it might offend some users? Is that scientific?

[mclane]

exactly john. its about science, not about personal things.
if we want to relate different results, we cannot do this without knowing the details.

[mclane]

Code: Select all

    Motor                      Punkte HiRyToGlLoChNaToRuNoBaChCoPrZaNo    S-B
01: Hiarcs11.1UCI              13,5/15 · = 1 1 1 = 1 = 1 1 1 1 1 1 1 1   92,00
02: Rybka v2.3.lk.w32          11,0/15 = · 0 1 = 0 0 1 1 1 1 1 1 1 1 1   69,00
03: TogaII1.3x4 [default]      10,5/15 0 1 · 1 0 1 1 = 0 1 = 1 = 1 1 1   68,50
04: Glaurung1.2.1-32bit        10,0/15 0 0 0 · = = 1 1 1 1 = 1 = 1 1 1   60,25
05: LoopMP 12.32               9,5/15  0 = 1 = · = = = = = 1 = 1 1 = 1   61,50
06: program x                  9,5/15  = 1 0 = = · 0 0 0 1 1 1 1 1 1 1   58,00
07: Naum 2.0                   8,5/15  0 1 0 0 = 1 · 1 = = 1 0 1 0 1 1   54,75
08: TogaII 1.2.1a              8,5/15  = 0 = 0 = 1 0 · 1 = = = = 1 1 1   53,75
09: Ruffian2.1.0               8,5/15  0 0 1 0 = 1 = 0 · = = 1 = 1 1 1   51,00
10: Now0704                    6,5/15  0 0 0 0 = 0 = = = · 1 1 0 1 1 =   36,75
11: Baron1.8.1.Uci             6,0/15  0 0 = = 0 0 0 = = 0 · 1 1 0 1 1   32,25
12: program y                  6,0/15  0 0 0 0 = 0 1 = 0 0 0 · 1 1 1 1   29,50
13: Colossus2006f              4,5/15  0 0 = = 0 0 0 = = 1 0 0 · = 0 1   28,00
14: ProDeo 1.2                 4,5/15  0 0 0 0 0 0 1 0 0 0 1 0 = · 1 1   19,75
15: Zarkov 4.5t                2,5/15  0 0 0 0 = 0 0 0 0 0 0 0 1 0 · 1    9,75
16: Now0604                    0,5/15  0 0 0 0 0 0 0 0 0 = 0 0 0 0 0 ·    3,25

120 Partien von 240 gespielt

a typical tournament with toga1.2.1 and toga1.3x4.
hardware AMD Athlon 2400+, blitz 10+10
as you can see toga1.2.1 has 2 points less.
Whatever i do, which tournament i run, the new toga is ahead of the old.

[IWB]

Hello Thorsten

I think you are joking, are'nt you?

16 games - one round vs each engine? You can nearly dice that result! Make a Elolist out of that and have a look on the possible deviation.

For some reason (and you know why) I can't reveal my complete list but I have 800 games vs 8 Top engines. 100 games each engine, changing colors, all of these engines played each other at least the same amount of games!

Toga II 1.2.1a : 2756 10 10 3100 51.0 % 2749 36.1 %
Toga II 1.3x4 : 2726 19 19 800 45.4 % 2758 41.1 %

As you can see they are 30 Elo away now and have only 10 and 19 point deviation - that means that Toga II 1.2.1a is for sure better the 1.3x4 (at least with my conditions (which arent too far away from yours)).

Bye
Ingo

Code: Select all

    Motor                      Punkte HiRyToGlLoChNaToRuNoBaChCoPrZaNo    S-B
01: Hiarcs11.1UCI              13,5/15 · = 1 1 1 = 1 = 1 1 1 1 1 1 1 1   92,00
02: Rybka v2.3.lk.w32          11,0/15 = · 0 1 = 0 0 1 1 1 1 1 1 1 1 1   69,00
03: TogaII1.3x4 [default]      10,5/15 0 1 · 1 0 1 1 = 0 1 = 1 = 1 1 1   68,50
04: Glaurung1.2.1-32bit        10,0/15 0 0 0 · = = 1 1 1 1 = 1 = 1 1 1   60,25
05: LoopMP 12.32               9,5/15  0 = 1 = · = = = = = 1 = 1 1 = 1   61,50
06: program x                  9,5/15  = 1 0 = = · 0 0 0 1 1 1 1 1 1 1   58,00
07: Naum 2.0                   8,5/15  0 1 0 0 = 1 · 1 = = 1 0 1 0 1 1   54,75
08: TogaII 1.2.1a              8,5/15  = 0 = 0 = 1 0 · 1 = = = = 1 1 1   53,75
09: Ruffian2.1.0               8,5/15  0 0 1 0 = 1 = 0 · = = 1 = 1 1 1   51,00
10: Now0704                    6,5/15  0 0 0 0 = 0 = = = · 1 1 0 1 1 =   36,75
11: Baron1.8.1.Uci             6,0/15  0 0 = = 0 0 0 = = 0 · 1 1 0 1 1   32,25
12: program y                  6,0/15  0 0 0 0 = 0 1 = 0 0 0 · 1 1 1 1   29,50
13: Colossus2006f              4,5/15  0 0 = = 0 0 0 = = 1 0 0 · = 0 1   28,00
14: ProDeo 1.2                 4,5/15  0 0 0 0 0 0 1 0 0 0 1 0 = · 1 1   19,75
15: Zarkov 4.5t                2,5/15  0 0 0 0 = 0 0 0 0 0 0 0 1 0 · 1    9,75
16: Now0604                    0,5/15  0 0 0 0 0 0 0 0 0 = 0 0 0 0 0 ·    3,25

120 Partien von 240 gespielt

a typical tournament with toga1.2.1 and toga1.3x4.
hardware AMD Athlon 2400+, blitz 10+10
as you can see toga1.2.1 has 2 points less.
Whatever i do, which tournament i run, the new toga is ahead of the old.

[Uri Blass]

Hello Thorsten

I think you are joking, are'nt you?

16 games - one round vs each engine? You can nearly dice that result! Make a Elolist out of that and have a look on the possible deviation.

For some reason (and you know why) I can't reveal my complete list but I have 800 games vs 8 Top engines. 100 games each engine, changing colors, all of these engines played each other at least the same amount of games!

Toga II 1.2.1a : 2756 10 10 3100 51.0 % 2749 36.1 %
Toga II 1.3x4 : 2726 19 19 800 45.4 % 2758 41.1 %

As you can see they are 30 Elo away now and have only 10 and 19 point deviation - that means that Toga II 1.2.1a is for sure better the 1.3x4 (at least with my conditions (which arent too far away from yours)).

Bye
Ingo

Code: Select all

    Motor                      Punkte HiRyToGlLoChNaToRuNoBaChCoPrZaNo    S-B
01: Hiarcs11.1UCI              13,5/15 · = 1 1 1 = 1 = 1 1 1 1 1 1 1 1   92,00
02: Rybka v2.3.lk.w32          11,0/15 = · 0 1 = 0 0 1 1 1 1 1 1 1 1 1   69,00
03: TogaII1.3x4 [default]      10,5/15 0 1 · 1 0 1 1 = 0 1 = 1 = 1 1 1   68,50
04: Glaurung1.2.1-32bit        10,0/15 0 0 0 · = = 1 1 1 1 = 1 = 1 1 1   60,25
05: LoopMP 12.32               9,5/15  0 = 1 = · = = = = = 1 = 1 1 = 1   61,50
06: program x                  9,5/15  = 1 0 = = · 0 0 0 1 1 1 1 1 1 1   58,00
07: Naum 2.0                   8,5/15  0 1 0 0 = 1 · 1 = = 1 0 1 0 1 1   54,75
08: TogaII 1.2.1a              8,5/15  = 0 = 0 = 1 0 · 1 = = = = 1 1 1   53,75
09: Ruffian2.1.0               8,5/15  0 0 1 0 = 1 = 0 · = = 1 = 1 1 1   51,00
10: Now0704                    6,5/15  0 0 0 0 = 0 = = = · 1 1 0 1 1 =   36,75
11: Baron1.8.1.Uci             6,0/15  0 0 = = 0 0 0 = = 0 · 1 1 0 1 1   32,25
12: program y                  6,0/15  0 0 0 0 = 0 1 = 0 0 0 · 1 1 1 1   29,50
13: Colossus2006f              4,5/15  0 0 = = 0 0 0 = = 1 0 0 · = 0 1   28,00
14: ProDeo 1.2                 4,5/15  0 0 0 0 0 0 1 0 0 0 1 0 = · 1 1   19,75
15: Zarkov 4.5t                2,5/15  0 0 0 0 = 0 0 0 0 0 0 0 1 0 · 1    9,75
16: Now0604                    0,5/15  0 0 0 0 0 0 0 0 0 = 0 0 0 0 0 ·    3,25

120 Partien von 240 gespielt

a typical tournament with toga1.2.1 and toga1.3x4.
hardware AMD Athlon 2400+, blitz 10+10
as you can see toga1.2.1 has 2 points less.
Whatever i do, which tournament i run, the new toga is ahead of the old.

Thorsten said that it is one of many tournaments that he has.
A possible explanation for different results is that things may be dependent on the opening that is used.

I do not know the reason for the difference.

Uri

[Tony Thomas]

How well is the new version of Shredder doing Ingo?

[mclane]

maybe you don't understand.
it happens in ANY tournament i do.

what you test if you run matches is mainly learning and book influence.

[IWB]

Thorsten said that it is one of many tournaments that he has.
...

Uri

What is the value of 100 diced tourneys?

Bye
Ingo